Three-dimensional human machine interface design is a critical component of any virtual environment (VE) application. Different from normal 2D user interface, the 3D user interface is more close to human intuition. However, people often find it inherently difficult to understand 3-D spaces and to perform actions in free space. Although we live and act in a 3-D world, the physical world contains many more cues for understanding and constraints and affordances for action that cannot currently be represented accurately in a computer simulation. Conventional human machine interface operation is performed by using keyboard, mouse, touch panel and remote controller. However, these four operation schemes have some disadvantages: it is convenient to use keyboard for data typing, but not for visual object control; the user must hold a device for operation while using a mouse or a remote controller; while operating the touch panel, the user is limited in a space where his/her finger can reach the touch panel.